key: cord-034690-x8lkngra
authors: Ridge, Carole A; Desai, Sujal R; Jeyin, Nidhish; Mahon, Ciara; Lother, Dione L; Mirsadraee, Saeed; Semple, Tom; Price, Susanna; Bleakley, Caroline; Arachchillage, Deepa J; Shaw, Elizabeth; Patel, Brijesh V; Padley, Simon PG; Devaraj, Anand
title: Dual-Energy CT Pulmonary Angiography (DECTPA) Quantifies Vasculopathy in Severe COVID-19 Pneumonia
date: 2020-10-29
journal: Radiol Cardiothorac Imaging
DOI: 10.1148/ryct.2020200428
sha: 
doc_id: 34690
cord_uid: x8lkngra

BACKGROUND: The role of dual energy computed tomographic pulmonary angiography (DECTPA) in revealing vasculopathy in coronavirus disease 2019 (COVID-19) has not been fully explored. PURPOSE: To evaluate the relationship between DECTPA and disease duration, right ventricular dysfunction (RVD), lung compliance, D-dimer and obstruction index in COVID-19 pneumonia. MATERIALS AND METHODS: This institutional review board approved this retrospective study, and waived the informed consent requirement. Between March-May 2020, 27 consecutive ventilated patients with severe COVID-19 pneumonia underwent DECTPA to diagnose pulmonary thrombus (PT); 11 underwent surveillance DECTPA 14 ±11.6 days later. Qualitative and quantitative analysis of perfused blood volume (PBV) maps recorded: i) perfusion defect ‘pattern’ (wedge-shaped, mottled or amorphous), ii) presence of PT and CT obstruction index (CTOI) and iii) PBV relative to pulmonary artery enhancement (PBV/PAenh); PBV/PAenh was also compared with seven healthy volunteers and correlated with D-Dimer and CTOI. RESULTS: Amorphous (n=21), mottled (n=4), and wedge-shaped (n=2) perfusion defects were observed (M=20; mean age=56 ±8.7 years). Mean extent of perfusion defects=36.1%±17.2. Acute PT was present in 11/27(40.7%) patients. Only wedge-shaped defects corresponded with PT (2/27, 7.4%). Mean CTOI was 2.6±5.4 out of 40. PBV/PAenh (18.2 ±4.2%) was lower than in healthy volunteers (27 ±13.9%, p = 0.002). PBV/PAenh correlated with disease duration (β = 0.13, p = 0.04), and inversely correlated with RVD (β = -7.2, p = 0.001), persisting after controlling for confounders. There were no linkages between PBV/PAenh and D-dimer or CTOI. CONCLUSION: Perfusion defects and decreased PBV/PAenh are prevalent in severe COVID-19 pneumonia. PBV/PAenh correlates with disease duration and inversely correlates with RVD. PBV/PAenh may be an important marker of vasculopathy in severe COVID-19 pneumonia even in the absence of arterial thrombus.

. The pathophysiology and imaging features of severe COVID-19 pneumonia have been the focus of considerable interest from the outset of the pandemic. In early disease, widespread ground-glass opacification predominates on thoracic computed tomography (CT) (2) (3) (4) (5) (6) , and is supposedly associated with highly compliant lungs and disrupted vasoregulation, (7) . Vascular dysregulation is believed to be consequence of exaggerated activation of inflammatory and coagulation cascades (termed 'immunothrombosis') (1, (8) (9) (10) (11) (12) . Later in the course of disease, CT more commonly shows consolidation and fibrosis associated with lower lung compliance (13) .

There is growing evidence from radiologic and pathologic studies of a significant vasculopathy in COVID-19 pneumonia (14) (15) (16) (17) : in a recent study of post-mortem lungs in COVID-19, there were widespread microthromboses and striking new vessel formation (16) .

Furthermore, based on qualitative analyses, a number of studies have highlighted the potential role of dual energy computed tomography pulmonary angiography (DECTPA) (15, (18) (19) (20) (21) . Accordingly, in the present study we aimed to evaluate the relationships between a quantitative measure of perfusion on DECTPA (relative perfused blood volume, PBV/PAenh) (22) , and i) disease duration, ii) right ventricular dysfunction on echocardiography iii) Ddimer levels and (iv) obstruction score (23) in patients with severe COVID-19 pneumonia. A secondary aim was to compare PBV/PAenh in COVID-19 pneumonia to that of normal volunteers.

This retrospective, observational study was approved by United Kingdom Health Research Authority (HRA), through the integrated research application system (IRAS # 265630); the need for informed consent was waived. No industry financial support was given to the authors, who had control of the data and the information submitted for publication at all times. Twenty seven consecutive patients (20 men, mean age 53.4 years (range, 24-70); 7 women, mean age 60.4 years (range, 54-66) were transferred to the Royal Brompton Hospital, United Kingdom -a national tertiary referral centre for patients with severe acute respiratory failure (SARF) -between March 21 st and May 24 th , 2020. Included patients were i) mechanically ventilated and ii) imaged with DECTPA to diagnose potential pulmonary thrombus (PT) (Figure 1 ). Patients on extracorporeal membrane oxygenation (ECMO) support were excluded given the potential confounding factor of contrast egress through the return ECMO cannula. Patients with a body mass index greater than 35kg/m 2 , and those with an inability to arm-raise above the head (to mitigate beam-hardening artifact) were excluded. Qualitative DETCPA findings in a subset of 18 patients with COVID-19 pneumonia have been previously reported (15) . This early data demonstrated perfusion defects on perfused blood volume (PBV) maps, when available, as well as the presence of vascular dilatation on conventional CT, hypercoagulability and an increased dead space in 39 patients suggesting that pulmonary angiopathy was accountable for hypoxia observed in patients requiring mechanical ventilation for severe COVID-19 pneumonia. This provided the impetus for the further quantitative study.

The diagnosis of severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) infection was confirmed on real-time reverse transcriptase polymerase chain reaction assay in all patients. The date of symptom onset was documented from clinical notes. Treating physicians determined the type and frequency of investigations in accordance with standard institutional practice. The results of the following investigations, performed within a maximum of 48 hours prior to DECTPA, were recorded: i) D-dimer levels; and ii) right ventricular dysfunction on transthoracic echocardiography, recorded as a binary variable of 'normal' or 'abnormal' with the latter graded as mild, moderate or severe right ventricular Healthineers, Forcheim, Germany). Images were acquired at a trigger of 120 HU using bolus tracking at the main pulmonary artery (PA), when available (n=26 DECTPA studies); to document contrast dynamics in each patient, the rate of PA enhancement was recorded in HU/second using bolus tracking CT images. The iodine content of each voxel was derived using a three-material-decomposition algorithm for air, soft-tissue, and iodine (23) and reconstructed using a "dense lung" PBV map with a maximum threshold of -200 Hounsfield units (HU). The upper threshold of -200 HU was identified as a cutoff, which included the typical ground glass opacity seen in viral pneumonia. This threshold applied only to qualitatively assessed PBV images, and not quantitative measurements, which included both aerated and non-aerated lung.

Qualitative DECTPA Analysis PBV images were reviewed in consensus by two thoracic radiologists (CAR and SPGP) with 14-and 30-years' experience, respectively, blinded to all clinical data using dual energy processing software (Syngo Via Dual Energy, Siemens Healthineers, Forcheim, Germany). A perfusion defect was defined as a segmental (25) region of hypoenhancement using an automatically generated colour scale. The proportion of aerated lung affected by perfusion defects was visually estimated as a percentage of both lungs. In cases of uncertainty, readers had the ability to measure PBV/PAenh, a threshold of < 20% was categorised as representing hypoperfusion (26) . The 'morphology' of perfusion defects was categorised as i) wedge-shaped (analogous to the appearance seen in patients with acute pulmonary embolism (PE) (27) , ii) mottled (as seen in idiopathic or chronic thromboembolic pulmonary hypertension (28) or iii) amorphous ( Figure 2 ). The pulmonary arterial tree was assessed to determine the presence of PT, the order of involvement (the main PA representing the "first order" artery) and if the predominant perfusion defect pattern was attributable to PT in a corresponding anatomical distribution. Of the 27 patients studied, 20 were intubated prior to transfer from another institution at an average of 2.8 ±3.9 days prior to admission whereas seven were intubated and mechanically-ventilated following admission to our institution. Four patients were selfventilating through a tracheostomy at the time of surveillance CT, and therefore ventilatory measurements were not available. Mean dynamic compliance was reduced (32.9±19.1 ml/cmH2O) and abnormal at the time of CT in 29/34 (85.3%). Mean PEEP was 10±2.7cm H20.

Arterial filling defects, consistent with PT, were present on ten initial and six surveillance DECTPA, with only one new PT identified on surveillance DECTPA. Thus, overall, PT was present in 11 patients; thromboses were seen in second order (n=2), third order (n=2), fourth order (n=3) and fifth order (n=4) arteries.

Visually-scored perfusion defects ( Figure 2 ) were present in 38/38 (100%) DECTPA studies, with a mean extent of 36.1±17.2% (Table 3) . At initial CT, perfusion defects were classified as wedge-shaped (n = 2), mottled (n = 4) and amorphous (n =21) ( The clinical and laboratory findings at the time of initial DECTPA are shown in Table   2 . Right ventricular function was abnormal in 33/39 (84.7%) patients and graded mild (n= 13), moderate (n = 12) and severe (n=8). There was no relationship between the rate of enhancement (mean 20±9HU/s) in the main PA and right ventricular dysfunction (p=0.46). The mean D-dimer level, abnormal in all patients, was 5500.05±8691.15 ng/ml. On linear regression analysis, there was an inverse relationship between PBV/PAenh on initial DECTPA and right ventricular dysfunction (β = -7.2, SE 1.9, p = 0.001), which remained significant after controlling for age, gender, BMI and ethnicity (β = -5.6, SE 1.6, p = 0.001). When RV dysfunction was subdivided into mild, moderate and severe grades, the statistical significance of this relationship was not maintained, potentially due to small sample size.

There was no statistically significant relationship between D-dimer levels or Qanadli (7). As further support for this 'vasocentric' hypothesis, we have shown that the morphology of perfusion defects was categorized as amorphous in the majority, as opposed to the more typical wedgeshaped peripheral defects seen in acute pulmonary embolism; the amorphous pattern is akin to that observed in idiopathic pulmonary hypertension and chronic thromboembolic pulmonary hypertension (15) .

The relationship between PBV/PAenh and disease duration may provide insights into the pathophysiology of COVID-19-related severe respiratory failure. It is hypothesised in the literature that in 'early' COVID-19-induced ARDS, compliance is preserved in the face of poor oxygenation and there is a poor response to high PEEP (7). This has been dubbed the 'type L' physiology, characterized morphologically by a pattern of ground-glass opacification on CT. By contrast, in those with progressive disease, there is then a transition to a physiology that is more in keeping with typical ARDS, namely with lower compliance and higher response to PEEP, termed the 'type H' phenotype. Our observations provide some support for this theory: PBV/PAenh was seen to improve with time in our patients, whereas compliance, although not likely to be related to vasculopathy, decreased with time from I n p r e s s disease onset. This seemingly neat categorization of COVID-19 lung physiology has obvious attractions. However, it seems highly likely that, in individual patients, there will be considerable overlap between the proposed phenotypes.

The observation of decreased compliance over time, may raise the question that it is volume loss due to fibrosis, and not improving perfusion that has given rise to an increased PBV/PAenh in patients imaged later in their disease course, however, we did not observe an increase in iodine density (i.e. concentration) with increasing time from symptom onset, but rather an increase in the relative iodine distribution in the lungs, corrected for opacification of the pulmonary artery (PBV/PAenh). This supports our belief that it is impaired pulmonary perfusion at the outset of the disease, and not loss of volume later in the disease, that is responsible for our observations.

In light of our findings, a brief discussion of right ventricular (RV) dysfunction in COVID-19 is warranted. In one recent study of 100 consecutive patients with COVID-19, there was echocardiographic evidence of RV dysfunction in 32% (31); RV dysfunction was also the most prevalent feature at follow-up. Importantly left ventricular function was preserved in the majority. The progression to right heart failure can manifest as cardiomegaly and right ventricular dilatation in the absence of myocarditis at autopsy in COVID-19 infection (17) . No patients in our cohort were diagnosed with myocarditis, leading us to believe that acute cor pulmonale was the most likely culprit (32) . Quantitative analysis of DECTPA in our cohort showed that decreasing PBV/PAenh, indicating poorer pulmonary perfusion, is linked with the presence of RV dysfunction, independent of age, sex, BMI or ethnicity. Of note, the rate of enhancement on dynamic bolus-tracking CT images of the PA did not correlate with right ventricular function in this cohort, allaying concerns that the observed decreased PBV/PAenh is simply an artifact of a blunted enhancement curve, as has been observed in pulmonary thromboembolism (22) . We believe this is an important finding suggesting, at least in part, that pulmonary vasculopathy may have a direct bearing on cardiac function in COVID-19.

There are obvious limitations in our study. The cohort size is small, and the retrospective methodology introduces potential biases. Another issue is that, in the absence of pathologic proof, causal arguments regarding the true meaning of the radiologic observations must be made with a little caution. However, there is growing consensus of significant vascular pathology in COVID-19 and, given that biopsy proof of vascular involvement is impracticable in the majority, we propose that non-invasive imaging biomarkers, such as PBV/PAenh, may be the best indirect marker of vasculopathy. This technique has also been validated by direct comparison with pulmonary blood flow measurement in swine models in the setting of both balloon arterial occlusion and in acute lung injury (33, 34) .

A further limitation is the lack of a comparable control group with severe acute respiratory failure without COVID-19 pneumonia. In fact, the preponderance of multiorgan infarction and thrombosis has also been observed in severe acute respiratory syndrome coronavirus (SARS-C0V) (35, 36) . This provides the impetus for future study of the utility of DECT in other pulmonary infections. An important technical limitation was that obese patients with a BMI greater than 35 kg/m 2 were not included in this study, due to concerns of image 'noise' and artifactual hypo-enhancement because of contrast egress through the return cannula, respectively; the presence of image noise is particularly problematic in the reconstruction of PBV images. Whilst this might be overcome by increasing exposure factors, it comes with the obvious penalty of an increasing radiation burden. For other technical reasons, patients on ECMO support were also excluded from our analyses. However, in the absence of 'normal' perfusion values for these patients, it was felt that their inclusion would not represent a like-for-like comparison of pulmonary enhancement. The presence of consolidated and atelectatic lung presented a challenge to analysis. This potential confounder was addressed by stipulating that PBV map visual analysis should exclude such portions of lung, as the presence of volume loss may impact the reader's assessment of regional iodine distribution. However, this consolidated or atelectatic lung was included in quantitative analysis, the rationale for this inclusion was that a quantitative comparison of pulmonary iodine over a large ROI compared to pulmonary arterial iodine (PBV/PAenh) was more likely to be accurate than purely visual analysis. A final consideration is that the majority of patients (21/27, 77 .8%) in our cohort were therapeutically anticoagulated I n p r e s s (Figure 4 ), and two of these patients received systemic thrombolysis with the remaining patients receiving prophylactic low molecular weight heparin. As a confounder, this might explain the higher PBV/PAenh later in the course of disease. As future randomized controlled trials are devised, we suggest that DECTPA is used as a surveillance tool, to determine whether the temporal pattern of pulmonary perfusion represents the natural history of the disease, or an anticoagulation-related effect.

In summary, perfusion defects on DECTPA are prevalent in severe COVID-19 pneumonia even in the absence of pulmonary thrombosis; PBV/PAenh is significantly I n p r e s s Bolus tracking was used with a region of interest (ROI) on the main pulmonary artery.

All patients were imaged with their arms extended cranially in a caudocranial injection; 3 patients were imaged in the prone position to alleviate oxygen requirements. Dose length product (DLP) was recorded and effective dose was obtained by multiplying the mean DLP value by the adult chest k-factor of 0.014 mSv*mGy−1*cm−1.

Two observers visually inspected for the following CT features:

i) Presence/absence of pulmonary arterial filling defect

ii)

The 'highest' order of PE (1, main pulmonary artery; 2, right or left ; 3, lobar; 4, segmental and 5, subsegmental)

The incidence of venous thrombosis was documented using compression ultrasound and/or CT venography reports.

Clinical predictors of mortality due to COVID-19 based on an analysis of data of 150 patients from Wuhan, China

Coronavirus Disease 2019 (COVID-19): Role of Chest CT in Diagnosis and Management

Pulmonary High-Resolution Computed Tomography (HRCT) Findings of Patients with Early-Stage Coronavirus Disease 2019 (COVID-19) in Hangzhou, China

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Chest CT Findings in Coronavirus Disease-19 (COVID-19): Relationship to Duration of Infection

Initial CT findings and temporal changes in patients with the novel coronavirus pneumonia (2019-nCoV): a study of 63 patients in Wuhan, China

COVID-19 pneumonia: different respiratory treatments for different phenotypes?

Incidence of thrombotic complications in critically ill ICU patients with COVID-19

Clinical and immunological features of severe and moderate coronavirus disease 2019

Abnormal coagulation parameters are associated with poor prognosis in patients with novel coronavirus pneumonia

Acute pulmonary embolism and COVID-19 pneumonia: a random association?

COVID-19 Complicated by Acute Pulmonary Embolism

Ventilatory Ratio in Hypercapnic Mechanically Ventilated Patients with COVID-19-associated Acute Respiratory Distress Syndrome

Hypoxaemia related to COVID-19: vascular and perfusion abnormalities on dual-energy CT

Pulmonary angiopathy in severe Covid-19: physiologic, imaging and hematologic observations

Pulmonary Vascular Endothelialitis, Thrombosis, and Angiogenesis in Covid-19

Pulmonary and cardiac pathology in African American patients with COVID-19: an autopsy series from New Orleans

Pulmonary infarction secondary to pulmonary thromboembolism in COVID-19 diagnosed with dual-energy CT pulmonary angiography

Radiology: Pulmonary Vascular Manifestations of COVID-19 Pneumonia

Diagnosis, Prevention, and Treatment of Thromboembolic Complications in COVID-19: Report of the National Institute for Public Health of the Netherlands

Pulmonary Embolism on CTPA in COVID-19 Patients

Clinical Significance of Late Phase of Lung Perfusion Blood Volume (Lung Perfusion Blood Volume) Quantified by Dual-Energy Computed Tomography in Patients With Pulmonary Thromboembolism

New CT index to quantify arterial obstruction in pulmonary embolism: comparison with angiographic index and echocardiography

Correlation analysis of dual-energy CT iodine maps with quantitative pulmonary perfusion MRI

Segmental Anatomy of the Lungs

Dualenergy CT angiography for assessment of regional pulmonary perfusion in patients with chronic thromboembolic pulmonary hypertension: initial experience

Dual-energy CT for imaging of pulmonary hypertension: challenges and opportunities

Dual-energy CT for the assessment of contrast material distribution in the pulmonary parenchyma

Pulmonary vasospasm in systemic sclerosis: noninvasive techniques for detection

The Spectrum of Cardiac Manifestations in Coronavirus Disease 2019 (COVID-19) -a Systematic Echocardiographic Study

Acute cor pulmonale in ARDS: rationale for protecting the right ventricle

Pulmonary perfused blood volume with dual-energy CT as surrogate for pulmonary perfusion assessed with dynamic multidetector CT

Quantitative Dual-Energy Computed Tomography Predicts Regional Perfusion Heterogeneity in a Model of Acute Lung Injury

Analysis of deaths during the severe acute respiratory syndrome (SARS) epidemic in Singapore: challenges in determining a SARS diagnosis